BUCHI NIRCal - Manual (Page 153)

Chemometrics
NIRCal 5.5 Manual, Version A 153
Result of the calculation
Row Name Description
59 Blow Up Parameter Title
60 Residual Blow Up Factor for residual limit calculation (default: 2)
61 Score Blow Up Factor for score limit calculation (default: 1.05)
63 Max C-Set Spectra Residual Highest value of the C-Set residuals
64 Max Allowed Residual for
Calibration Residual limit of the calibration for the application (max.* 2)
Q-Value protocol
67 Q-Value V5 Q-Value of the calibration according version 5. The higher
the Q-Value the best the calibration.
68 Validation Method Selected Validation method: Validation Set or Cross
Validation
70 C-Set Residual too big* Number of C-Set spectra with too high residual = rejection of
known (theoretically: 0 but always tested!)
71 V-Set Residual too big* Number of V-Set spectra with too high residual = rejection of
unknown (should be zero)
73 Num Properties Number of properties in the calibration (should be always 1
for the application)
74 Rel. Consistency* The C- and V-Set should have similar low standard error of
prediction (BIAS corrected):
Abs(SEC-SEP)/(Abs(SEP)+1). Should be close to 0
75 Weighted BIAS* The absolute V-Set BIAS should be as close to zero as
possible:
Abs(Vset-BIAS)/Abs(Range)
76 Validity* The V-Set regression coefficient should be near to 1:
1-VsetRegr. Validity should be close to 0
77 Comparability* C- and V-Set should have similar high regression coefficient
of prediction:
Abs(CsetRegr-VsetRegr). Comparability should be close to 0
78 Precision* V-Set standard error of prediction (BIAS corrected) should be
low:
SEP/Abs (Range)
79 Weighted Accuracy* V-Set residual sum of squares=RSS should be low (not BIAS
corrected):
Abs(Sum((Orig-Predicted)^2)/Abs (Range)
* used with different weighting values.